| 電子情報通信学会 研究会発表申込システム 講演論文 詳細 |
技報閲覧サービス [ログイン] 技報アーカイブ |
| トップに戻る | 前のページに戻る | [Japanese] / [English] |
| 講演抄録／キーワード | ||
| 講演名 | 2008-03-20 15:15 ［ポスター講演］Robust noise suppression algorithm using the only Kalman filter theory for white and colored noises ○Nari Tanabe（Tokyo Univ. of Science, Suwa）・Toshihiro Furukawa（Tokyo Univ. of Science）・Shigeo Tsujii（Inst. of Information Security）  SP2007-195 |
|
| 抄録 | （和） | This paper presents a noise suppression algorithm using only the Kalman filter theory with canonical state space models: (i) a state equation is composed of the speech signal, and (ii) an observation equation is composed of the speech signal and additive noise. The algorithm aims to achieve simple and robust noise suppression without the conception of the AR (autoregressive) system, while many conventional methods based on the Kalman filter usually performs the parameter estimation algorithm of AR system and then the Kalman filter algorithm. It should be noted that driving source is a colored signal (speech signal) in the proposed canonical state space models. As is known well, on the other hand, the Kalman filter theory is usually applied to the model in which the driving source is white signal. Therefore, in case that the Kalman filter theory is applied to the proposed canonical state space model, we must examine the effect that the colored driving source has on the estimation accuracy of the state variables. But, unfortunately, it is very difficult to analyze the effect of colored driving source to the estimation accuracy of the state theoretically as is described later in detail. This paper shows, by some numerical simulations, that the Kalman filter algorithm applied to the proposed canonical state space models functions well. |
| （英） | This paper presents a noise suppression algorithm using only the Kalman filter theory with canonical state space models: (i) a state equation is composed of the speech signal, and (ii) an observation equation is composed of the speech signal and additive noise. The algorithm aims to achieve simple and robust noise suppression without the conception of the AR (autoregressive) system, while many conventional methods based on the Kalman filter usually performs the parameter estimation algorithm of AR system and then the Kalman filter algorithm. It should be noted that driving source is a colored signal (speech signal) in the proposed canonical state space models. As is known well, on the other hand, the Kalman filter theory is usually applied to the model in which the driving source is white signal. Therefore, in case that the Kalman filter theory is applied to the proposed canonical state space model, we must examine the effect that the colored driving source has on the estimation accuracy of the state variables. But, unfortunately, it is very difficult to analyze the effect of colored driving source to the estimation accuracy of the state theoretically as is described later in detail. This paper shows, by some numerical simulations, that the Kalman filter algorithm applied to the proposed canonical state space models functions well. | |
| キーワード | （和） | robust noise suppression / Kalman filter theory / canonical space state models / state equation / observation equation / white and colored noises / color driving source / high performance and quality |
| （英） | robust noise suppression / Kalman filter theory / canonical space state models / state equation / observation equation / white and colored noises / color driving source / high performance and quality | |
| 文献情報 | 信学技報, vol. 107, no. 551, SP2007-195, pp. 51-56, 2008年3月. | |
| 資料番号 | SP2007-195 | |
| 発行日 | 2008-03-13 (SP) | |
| ISSN | Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 | |
| 著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します．(許諾番号：10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) | |
| PDFダウンロード | SP2007-195 |
|
| 研究会情報 | |
| 研究会 | SP |
| 開催期間 | 2008-03-20 - 2008-03-21 |
| 開催地（和） | 東大 |
| 開催地（英） | Univ. Tokyo |
| テーマ（和） | 国際ワークショップ(20日），研究会(21日) 「聴覚・音声・言語とその障害，一般」 |
| テーマ（英） | International Workshop (Mar 20), Speech Production, Speech Perception, Hearing and Speech, etc. (Mar 21) |
| 講演論文情報の詳細 | |
| 申込み研究会 | SP |
| 会議コード | 2008-03-SP |
| 本文の言語 | 英語 |
| タイトル（和） | |
| サブタイトル（和） | |
| タイトル（英） | Robust noise suppression algorithm using the only Kalman filter theory for white and colored noises |
| サブタイトル（英） | |
| キーワード(1)（和/英） | robust noise suppression / robust noise suppression |
| キーワード(2)（和/英） | Kalman filter theory / Kalman filter theory |
| キーワード(3)（和/英） | canonical space state models / canonical space state models |
| キーワード(4)（和/英） | state equation / state equation |
| キーワード(5)（和/英） | observation equation / observation equation |
| キーワード(6)（和/英） | white and colored noises / white and colored noises |
| キーワード(7)（和/英） | color driving source / color driving source |
| キーワード(8)（和/英） | high performance and quality / high performance and quality |
| 第1著者 氏名（和/英/ヨミ） | Nari Tanabe / Nari Tanabe / |
| 第1著者 所属（和/英） | Tokyo University of Science, Suwa (略称： 諏訪東京理科大) Tokyo University of Science, Suwa (略称： Tokyo Univ. of Science, Suwa) |
| 第2著者 氏名（和/英/ヨミ） | Toshihiro Furukawa / Toshihiro Furukawa / |
| 第2著者 所属（和/英） | Tokyo University of Science (略称： 東京理科大) Tokyo University of Science (略称： Tokyo Univ. of Science) |
| 第3著者 氏名（和/英/ヨミ） | Shigeo Tsujii / Shigeo Tsujii / |
| 第3著者 所属（和/英） | Institute of Information Security (略称： 情報セキュリティ大) Institute of Information Security (略称： Inst. of Information Security) |
| 講演者 | 第1著者 |
| 発表日時 | 2008-03-20 15:15:00 |
| 発表時間 | 90分 |
| 申込先研究会 | SP |
| 資料番号 | SP2007-195 |
| 巻番号（vol） | vol.107 |
| 号番号（no） | no.551 |
| ページ範囲 | pp.51-56 |
| ページ数 | 6 |
| 発行日 | 2008-03-13 (SP) |
[研究会発表申込システムのトップページに戻る]